rakshas meant strong and powerful, but it soon took the sense of giant and barbarian, and in this sense it occurs in the Veda together with Yâtudhâna.

Another Vaidik epithet applied as it seems to wild tribes, infesting the seats of the Aryas, is "anagnitra," they who do not keep the fire. Thus we read," Agni, drive away from us the enemies,-tribes who heep no sacred fires came to attack us. Come again to the earth, sacred god with all the immortals, come to our libation."

The same races are called " Kravyâd," or flesh-eaters. In a famous hymn of Vasishtha we read: "Indra and Soma, burn the Rakshas, destroy them, throw them down, ye two Bulls, the people that grow in darkness. Hew down the madmen, suffocate them, kill them, hurl them away and slay the voracious."

"Indra and Soma, up together against the cursing demon! May he burn and hiss like an oblation in the fire! Put your everlasting hatred upon the villain, who hates the Brahman, who eats flesh, and whose look is abominable.

"Indra and Soma, hurl the evil-doer into the pit, into unfathomed darkness. May your strength be full of wrath to hold out, that no one may come out again."

Kravyád, flesh-eater, means people who eat raw meat, peopάyo, and they are also called âmâdas, wμopάyoɩ, or raw-eaters, for the cooking of meat was a distinguishing feature of civilized nations, and frequently invested with a sacrificial character. Agni, who in the Veda is the type of the sacrifice, and with it of civilization and social virtues, takes an entirely different character in his capacity of Kravyâd," or flesh-eater. He is represented under a form as hideous as the beings he is invoked to devour. He sharpens his two iron-tusks, puts the enemies into his mouth and swallows them. He heats the edges of his shafts, and sends them into the hearts of the Rakshas. He tears their skin, minces their members, and throws them before the wolves to be eaten by them or by the shrieking vultures. These Rakshas are themselves called "akitas," mad, and "mûradevâs," worshippers of mad gods. Nay they are even taunted with eating human flesh, and are called "asutripas," as enjoying the life of other men. In the Rigveda, we read, "The Yâtudhânas who gloat on the bloody flesh of men or

horses, and steal the milk of the cow, o Agni, cut off their heads with thy fiery sword."

All these epithets seem to apply to hostile, and most likely aboriginal races, but they are too general to allow us the inference of any ethnological conclusions. The Vaidik Rishis certainly distinguish between Arian and non-Arian enemies. The gods are praised for destroying enemies, Arian as well as barbarian (dâsâ ka vritrâ hatam, âryâni ka), and we frequently find the expression, "Kill our Arian enemies, and the Dâsa enemies, yea, kill all our enemies." But there is no allusion to any distinct physical features such as we find in later writings. The only expression that might be interpreted in this way is that of "susipra," as applied to Arian gods. It means "with a beautiful nose." As people are fain to transfer the qualities which they are most proud of in themselves, to their gods, and as they do not become aware of their own good qualities except by way of contrast, we might conclude that the beautiful nose of Indra was suggested by the flat-noses of the aboriginal races. Tribes with flat or with even no noses at all, are mentioned by Alexander's companions in India, and in the hymns of the Rigveda Manu is said to have conquered Vi-sisipra (Pada-text, visi-sipra), which may be translated by "nose-less." The Dâsa or barbarian is also called vrishasipra in the Veda, which seems to mean goat or bullnosed, and the "Anâsas" enemies whom Indra killed with his weapon (Rv. V, 29, 10), are probably meant for noseless (a-nâsas), not, as the commentator supposes, for faceless (an-âsas) people.

In the Brahmanas, which represent a new period of Vaidik literature, the Nishâdas occur under more distinct features. In the Aitareya-brâhmana, they are once mentioned in the same category with thieves and criminals, who attack men in forests, throw them into wells, and run away with their goods (Nishâdâ vâ, Selagà vâ, pâpakrito vâ).

In some of the later Brâhmanas also, the Pankavinsa, for instance, the Nishâdas occur, and we there find, that they now live not only in forests but in villages. But there also, they are distinct from the castes as well as from the great mass of the people, the latter, though not under Brahmanic discipline, being yet considered as of Arian origin. This latter class, the Vrâtyas, are de

scribed as differing from the Brahmanic laity in laws, customs, and pronunciation, but not in language. They could be readmitted into the Brahmanic community after performing certain rites and penances prescribed by law. Their name is Vrâtya, but never Nishâda In the Taittirîya-brâhmana, we find after the four castes (Brâhmana, Raganya, Vaisya and Sûdra), other names, such as Magadha, Sailûsha, Naishâda, Vrâtya, Kaivarta, Kirâta, Kândâla, etc., but again no description of their physical peculiarities.

This is very different in later works. In the Vishnu-purâna (page 100, ed. Wilson), the type of the Nishâda is given,-"a being of the complexion of a charred stake, with flattened features, and of dwarfish stature." The inhabitants of the Vindhya mountains are called his descendants. According to the Matsya-purâna, they were as black as collyrium. According to the Bhagavata-purâna, they had short arms and legs, were black as a crow, with projecting chin, broad and flat nose, red eyes, and tawny hair. The Padma-purâna adds a wide mouth, large ears, and a protuberant belly, and particularises their posterity as Kirâtas, Bhillas, Bahanakas, Bhramaras, and Pulindas.

From the most ancient times therefore to the period of the Purânas, we meet everywhere with indications, more or less distinct, of two races brought into contact in the Indian peninsula. A most vivid description of their physical peculiarities at the present time is given by Mr. Hodgson. In one of his articles published in the Journal of the Asiatic Society of Bengal (1849, p. 710), he writes:

"A practised eye will distinguish at a glance between the Arian and Tamulian (i. e. Nishâda) style of features and form-a practised pen will readily make the distinction felt-but to perceive and to make others perceive, by pen or pencil, the physical traits that separate each group or people of Arian or of Tamulian (Nishâda) extraction. from each other group, would be a task indeed! In the Arian form there is height, symmetry, lightness and flexibility: in the Arian face an oval contour with ample forehead and moderate jaws and mouth a round chin, perpendicular with the forehead, a regular set of distinct and fine features; a well-raised and unexpanded nose, with elliptic nares; a well-sized and freely opened eye, running directly across the face; no want of eye-brows, eye-lash, or beard; and lastly, a clear

brunet complexion; often not darker than that of the most southern Europeans.

"In the Tamulian (Nishâda) form, on the contrary, there is less height, less symmetry, more dumpiness and flesh in the Tamulian face, a somewhat lozenge contour caused by the large cheek bones; less perpendicularity in the features to the front, occasioned not so much by defect of forehead or chin, as by excess of jaws and mouth; a larger proportion of face to head, and less roundness in the latter; a broader, flatter face, with features less symmetrical, but perhaps more expression, at least of individuality; a shorter, wider nose, often clubbed at the end and furnished with round nostrils; eyes less, and less fully opened, and less evenly crossing the face by their line of aperture; ears larger; lips thicker; beard deficient; colour brunet as in the last, but darker on the whole, and, as in it, various. Such is the general description of the Indian Arians and Turanians."

In other places Mr. Hodgson undertakes indeed to give some characteristic marks by which the principal sub-divisions of this Non-Arian, or Nishâda, stock might be distinguished in different parts of India. But though they would suffice to indicate at once the Nishâda in the Dekhan or in the jungles of Gondvân, in the slopes of the Vindhya or in the valleys of the Brahmaputra, in the Tarai or in the Ghats of the Himalaya, from his Arian neighbour, they are hardly sufficient to separate the Tamulian proper from the Kol, the Kol from the Garo, the Garo from the Lepcha, the Lepcha from the Bhotîya. Mr. Hodgson also, admits, in several places, that, on the whole, there is but one stamp impressed on all the Aborigines of India, that will admit of scientific definition. This stamp, he says, is the Mongolian "Look steadfastly at any man of an aboriginal race (an ubiquitarian Dhanger for instance), and say if a Mongol origin is not palpably inscribed on his face".

SECOND SECTION.

Ethnology v. Phonology.

ETHNOLOGY, therefore, as a physical science, would hardly bring us beyond a general conviction that India is inhabited by two different races of men. Nor should we, in our phonological studies, either expect or desire more than general hints from physical ethnology, The proper and rational connection between these two sciences is that of mutual advice and suggestion, but nothing more. Much of the confusion of terms and indistinctness of principles, both in ethnology and phonology, are due to the combined study of these heterogeneous sciences. Ethnological race and phonological race are not commensurate, except in ante-historical times, or perhaps at the very dawn of history. With the migrations of tribes, their wars, their colonies, their conquests and alliances, which, if we may judge from the effects, must have been much more violent in the ethnic, than ever in the political periods of history, it is impossible to imagine that race and language should continue to run parallel. The physiologist should pursue his own science unconcerned about language. Let him see how far the skulls, or the hair, or the colour, or the skin of different tribes admit of classification; but to the sound of their words his ear should be as deaf as the ornithologist's to the notes of caged birds. If his Caucasian class includes nations or individuals speaking Arian (Greek), Turanian (Turk), and Semitic (Hebrew) languages, it is not his fault. His system must not be altered in order to suit another system. There is a better solution both for his difficulties and for those of the phonologist than mutual compromise. The phonologist should collect his evidence, arrange his classes, divide and combine, as if no Blumenbach had ever looked at skulls, as if no Camper had measured facial angles, as if no Owen had examined the basis of a cranium. His evidence is the evidence of language, and nothing else; this he must follow, even though it be in the teeth of history, physical or political. Would he scruple to call the language of England Teutonic, and class it with the Low